Search

Characterization of patterns of alcohol consumption during pregnancy encompasses multiple factors such as magnitude, frequency, and timing of exposure throughout gestation. Traditional statistical models are limited in dealing with multivariate and diverse patterns of exposure as in the context of this analysis. We propose a finite mixture model-based approach to derive clusters of alcohol exposure of participants in the Safe Passage Study (PASS). Daily alcohol consumption data for 11,083 pregnant women have been clustered in 10 different exposed groups. The resulting cluster analysis was able to characterize alcohol consumption in a comprehensive framework capable of taking into account both quantity and timing of exposure as well as the occurrence of binge drinking.